JF Avila-Tomas, E Olano-Espinosa, C Minué-Lorenzo, FJ Martinez-Suberbiola, B Matilla-Pardo, ME Serrano-Serrano, E Escortell-Mayor and Dej@loBot Team. Effectiveness of a chat-bot for the adult population to quit smoking: protocol of a pragmatic clinical trial in primary care (Dejal@). BMC Medical Informatics and Decision Making; 19(1): 1-10, 2019

1. STUDY PROTOCOL Open Access
Effectiveness of a chat-bot for the adult
population to quit smoking: protocol of a
pragmatic clinical trial in primary care
(Dejal@)
J. F. Avila-Tomas1,2*
, E. Olano-Espinosa3
, C. Minué-Lorenzo4
, F. J. Martinez-Suberbiola5
, B. Matilla-Pardo6
,
M. E. Serrano-Serrano7
, E. Escortell-Mayor8,9
and the Group Dej@lo
Abstract
Background: The wide scale and severity of consequences of tobacco use, benefits derived from cessation, low
rates of intervention by healthcare professionals, and new opportunities stemming from novel communications
technologies are the main factors motivating this project. Thus, the purpose of this study is to assess the
effectiveness of an intervention that helps people cease smoking and increase their nicotine abstinence rates in the
long term via a chat-bot, compared to usual practice, utilizing a chemical validation at 6 months.
Methods: Design: Randomized, controlled, multicentric, pragmatic clinical trial, with a 6-month follow-up. Setting:
Healthcare centers in the public healthcare system of the Community of Madrid (Madrid Regional Health Service).
Participants: Smokers > 18 years of age who attend a healthcare center and accept help to quit smoking in the
following month. N = 460 smokers (230 per arm) who will be recruited prior to randomization. Intervention group: use
of a chat-bot with evidence-based contents to help quit smoking. Control group: Usual treatment (according to the
protocol for tobacco cessation by the Madrid Regional Health Service Main variable: Continuous nicotine withdrawal
with chemical validation (carbon monoxide in exhaled air). Intention-to-treat analysis. Difference between groups in
continuous abstinence rates at 6 months with their corresponding 95% confidence interval. A logistic regression model
will be built to adjust for confounding factors. Results: First expected results in January 2020.
Discussion: Providing science-based evidence on the effectiveness of clinical interventions via information
technologies, without the physical presence of a professional, is essential. In addition to being more efficient, the
characteristics of these interventions can improve effectiveness, accessibility, and adherence to treatment. From an
ethics perspective, this new type of intervention must be backed by scientific evidence to circumvent pressures from
the market or particular interests, improve patient safety, and follow the standards of correct practices for clinical
interventions.
Trial registration: ClinicalTrials.gov, reference number NCT 03445507.
Keywords: Smoking, Tobacco cessation, Primary care, Cell phone use, Chat-bot, Dialog systems
© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
* Correspondence: joseavil@gmail.com
1
Universidad Rey Juan Carlos, Madrid, Spain
2
Healthcare center Sta. Isabel, Madrid Regional Health Service, 28911,
Leganés, Madrid, Spain
Full list of author information is available at the end of the article
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249
https://doi.org/10.1186/s12911-019-0972-z

2. Background
Tobacco use is the main cause of preventable morbidity
and mortality in the world, directly accounting for five
million deaths yearly [1]. The number of smokers who
attempt quitting on their own every year is high, but
only 2–3% remain abstinent after 12 months [2]. Health-
care professionals are very effective and efficient in their
interventions on the smoker, with over threefold success
rates of long-term abstinence. Combining behavioral and
pharmacological treatment yields the best results [2–4].
However, only 1 in 20 attempt to quit is supervised by
a professional [2]. In the public healthcare system of the
Community of Madrid (Madrid Regional Health Ser-
vice), 84% of smokers who attended an outpatient con-
sultation in 2008 did not receive advice for tobacco
cessation [5], a figure similar to those observed in other
countries [6, 7]. Factors related to these low intervention
rates have been identified, and these include lack of
training for healthcare professionals and their perception
that such interventions are not useful and they lack the
time to implement them [8].
A possible solution is developing effective, brief, and
simpler interventions, such as the “Very Brief Advice
“that consists of offering the patient help to quit smok-
ing regardless of their motivation. Many patients re-
spond positively to this proposal, even those who had
not considered making an attempt previously to the
offer [9]. Compared to the usual intervention, this re-
sults in over 50% more attempts to cease smoking [10].
However, to achieve this, a resource backed by scientific
evidence and easy to access must be available, such as
the tobacco cessation consultations in the British Na-
tional Health System (NHS).
The boom in information and communication tech-
nologies, like the internet or smartphones, open up new
therapeutic perspectives. In 2016, there were 7 billion
customers of mobile phone services, almost 97% of the
population worldwide. Mobile broadband has continu-
ously grown, with 84% of the world population using it
[11], and the smartphone has become the main and
most accessible personal computer in the majority of
countries. On the other hand, patients aspire to play a
greater role in their health management and increasingly
search for more information in the internet (Table 1).
The resulting opportunities create a new framework to
empower the patient and improve clinical outcomes and
health expenditure [12].
In terms of scientific evidence, a systematic review of
59 clinical trials on interventions to quit smoking via
smartphones [13] concluded that text messages double
the success rates of nicotine abstinence with biochemical
validation, showing a risk ratio (RR) of 2.16 (Confidence
interval -CI- 95%: 1.77–2.62). The relevant Cochrane
review [14] comprised 12 clinical trials that included
11,885 patients, with a combined RR = 1.7 (CI 95%:
1.46–1.90) for abstinence at 6 months, and a RR =
1.83 (CI 95%: 1.54–2.19) using the data of the six
clinical trials that chemically validated the abstinence,
both compared to usual practice. The evaluated trials
were based on interventions mainly using Short Mes-
sage Service (SMS), although one employed videos,
and all were conducted in high-income countries with
strong policies for tobacco control. They also did not
find clinical trials, published or in-progress, studying
the effectiveness of applications (app) for mobile
phones or tablets, despite their proliferation.
In a 2013 review about the adequacy of scientific evi-
dence for existing apps to aid people cease smoking in
the USA at the time, Abroms et al. [15] could not find
any that followed science-based recommendations by
clinical practice guidelines and warned about the poten-
tial negative effects on the population’s health.
In 2017, a new review [16] aimed to assess the scien-
tific content of the most used commercial apps in the
USA for helping to quit smoking as well as the ones
available. Only six apps were identified as being partially
science-based, of which three (50%) were available in at
least one app store, and just two of the top 50 recom-
mended apps in app stores (4%) had some scientific
basis, but it was not possible to differentiate them from
those not based on scientific evidence.
A chat-bot is not a software that needs to be installed
in a smartphone or tablet, it respects the privacy of the
patient scrupulously, and its learning curve is very short
[17]. This computer software has a conversation inter-
face that can both answer questions posed by the user in
Table 1 Characteristics of the web search for health information
Advantages Downsides
Ubiquity Unequal quality of information
Immediacy An identified source cannot always be found
Ability to access specific population segments The date when the content was updated cannot always be found
Automation of messages Need to adapt the message to the targeted population
Possibility of contact and learning among peers Risk of information overload for the patient
Translation from: Casado, S. El Papel de la Información en el empoderamiento del paciente en Basagoiti I. Alfabetización en salud. De la información a la acción
[pdf]. Valencia: ITACA/TSB; 2012. ISBN: 978–84- 695-5267-4 Available at http://www.salupedia.org/alfabetizacion/
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 2 of 10

3. a natural language (that commonly used by people with
all their variants) and ask them questions [18]. Although
it is not a new technology, the recent technical revolu-
tion for the interpretation of natural languages, together
with the above-mentioned advantages and certain down-
sides of apps, make chat-bots ideal tools for the pur-
poses of this intervention. Since chat-bots “understand”
requests expressed with the complexity and variability of
human language, they provide a component of techno-
logical “humanization” that other interfaces based on
menus and chat-buttons lack. This way, they are capable
of returning a personalized answer and add a component
of user loyalty and usability of the tool [18, 19].
The chat-bot to be evaluated in this trial has been spe-
cifically designed by experts in tobacco addiction and arti-
ficial intelligence incorporating gamification, cognitive-
behavioral, motivational, problem-solving, and relapse-
preventing components. Such components form an inte-
gral part of science-based interventions recommended by
clinical practice guidelines [3, 4].
We aim to assess whether the effectiveness of the
usual intervention in primary care can be improved [2]
by providing healthcare professionals with guidance to
help patients accept the offer to quit smoking [10], so
that their motivation to intervene on their patients will
increase by decreasing their workload. We also estimate
that this type of intervention will improve the accessibil-
ity of patients to an evidence-based treatment.
The wide scale and severity of consequences of
tobacco use, benefits derived from cessation, low rates of
intervention by healthcare professionals, and new oppor-
tunities stemming from novel communications tech-
nologies in addition to the absence of scientific evidence
on its effectiveness and the lack of such tools designed
by experts according to the guidelines of the evidence-
based clinical practice guidelines are the main factors
motivating this project.
Methods
Aims
Main objective
To assess the effectiveness of an intervention in primary
healthcare to help smokers cease their tobacco addiction
via a chat-bot for smartphones compared to usual prac-
tice, as measured through biochemical validation of con-
tinuous abstinence at 6 months.
Secondary objectives
1. To assess the intervention effectiveness compared
to usual practice as measured through self-reported
continuous abstinence at 6 months.
2. To assess the intervention effectiveness compared
to usual practice in terms of the improvement in
their quality of life measured by the EuroQol-5D-5
L questionnaire.
3. To evaluate the cost-utility ratio of the intervention
compared to usual practice in primary healthcare.
4. To assess the intervention effectiveness in terms of
improved adherence to pharmacological treatment
for tobacco cessation.
Design
The clinical trial will be pragmatic, randomized, con-
trolled, and multicentric. It will last one year, six months
for recruiting and six months of follow-up. A cost-utility
study will be conducted from the perspective of the
funder for a 1-year timeframe. (Figs. 1,2)
Setting
Primary healthcare centers in the Madrid Regional
Health Service
Participants
Patients > 18 years of age who smoke and attend their
primary care doctor or nurse consultation during the in-
clusion period.
Criteria for inclusion
1. Healthcare centers: Doctors and nurses from 34
healthcare centers in the Madrid Regional Health
Service who agreed to participate, amounting to a
total of 248 participants.
2. Patients: Being > 18 years of age; having smoked
more than one cigarette per day during the last
month; accepting the offered help to quit smoking
in the following month; owning a smartphone
where a messaging app can be installed; being
reachable during the six months following the
beginning of the intervention; agreeing to
participate and signing informed written consent.
The criteria for exclusion are: Significant communica-
tion barriers; Addiction to other substances; And partici-
pating in another dishabituation program or clinical trial
during the study period.
Sample size
The sample size was calculated based on the outcome of
a clinical trial recently conducted of usual clinical care
in Madrid public health system that employs a similar
definition of continuous abstinence as our study [5]. It
detected an abstinence rate of 9.6% at 6 months in the
control group. We estimate that the intervention to be
assessed will double this fig [13, 14]..
Based on these data, and considering a type I error of
5% and a power of 80%, the required number of patients
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 3 of 10

4. Fig. 1 Protocol
Fig. 2 Pat plot intervention Dej@lo
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 4 of 10

5. will be 418. After estimating 10% of losses to follow-up,
the required sample size is 460 smokers (230 in each
group).
Recruitment
Each healthcare professional will select patients who
smoke, attend the health center for any reason within
the recruitment period, and meet the inclusion criteria.
They will inform patients about the characteristics of the
trial and offer them to participate. Those who accept will
be asked to provide informed written consent. In case
they decline, their gender, age, and reason for declining
will be recorded.
Various strategies will be implemented in order to in-
crease adherence to the protocol by participating re-
searchers, such us personal follow-up upon achievement
of protocol objectives and acknowledgement of their ef-
fort via e-mail, and offers of certified training sessions.
Randomized allocation: Subjects will be allocated into the
intervention or control group by simple randomization
using a software installed in the data collection notebook
(DCN). This will be done following the inclusion of patients
and data collection in the DCN at the initial visit, which
guarantees the masking of the randomization sequence to
the professionals recruiting patients. The group assigned to
the chat-bot will be the intervention group and those
treated with usual care by their doctor or nurse will be the
control group.
Criteria for abandonment and withdrawal
Participants will be able to abandon the study at any
time and researchers will be able to withdraw any pa-
tient under the following circumstances: not meeting the
inclusion criteria; Onset of a severe illness during the
clinical trial; Or adverse event, and inability to comply
with the study requirements.
The number of abandonments and withdrawals, as
well as the reasons for them, will be recorded. With-
drawn patients will not be excluded from the intention-
to-treat analysis.
Masking
The design of the study does not allow for masking the
patient to the received treatment. However, this limita-
tion is compensated by the objective measure of the
main outcome variable (smoking abstinence) and the
random allocation of patients into study arms. Addition-
ally, researchers in charge of the statistical analysis will
be blinded to the identity of patients in each treatment
group.
Interventions
The intervention strategy to follow in both arms is based
on the five A’s of the clinical practice guideline by the
Agency for Healthcare Research and Quality [4]: ask, ad-
vise, assess, assist, and arrange. This way, all patients will
be questioned in person about their tobacco addiction, re-
ceive advice from their doctor or nurse to cease smoking,
and their willingness to quit will be evaluated. Patients
who accept the offered help for the month following con-
sultation will be randomly assigned to the intervention or
control group, where they will receive a personal interven-
tion combining behavioral and pharmacological treatment
based on scientific evidence [2–4]. The intervention will
be organized into several follow-up sessions, either on-line
via the chat-bot (in which case the patient can access the
intervention at their convenience) or face-to-face with
their healthcare professional in the healthcare center.
Intervention group
Patients assigned to the intervention group will use a
chat-bot whose script for the smoking cessation process
has been developed using contents based on current evi-
dence. Subjects will download a messaging application
on their smartphone for this purpose and will access the
chat-bot with an assigned personal password. The chat-
bot will guide them through all the stages of the dishabi-
tuation process. The chat is bidirectional, employs
multiple media formats, and provides patients with auto-
matic, science-based advice on cognitive-behavioral,
motivational, relapse-preventative, and problem-solving
techniques. This will take place with different periodicity
depending on the quit date and other characteristics of
the patient (personal choice, type of tobacco use, per-
sonal risk situation, prescribed medication, abstinence-
related symptoms, and the evolution of coping with
abstinence). The chat-bot also offers information about
useful medication to quit smoking prescribed by their
healthcare professional, and will recommend how to face
problems related to the cessation process through advice
and relaxation exercises available in different formats,
such as video, computer graphics, games, and web links.
Additionally, the chat-bot incorporates gamification ele-
ments (games for adults for acquiring new knowledge
and skills), including a system to earn scores and badges
that will provide access to specific information depend-
ing on the abstinence period and the needs of the
individual. Specific aspects of the game mechanics
(score, levels for different abstinence periods) and dy-
namics (rewards, acknowledgment) will be developed.
Contact with the patient will begin at day − 15
, where the
above-mentioned strategies will be established, and daily
interactive appointments will be set until the quit date.
From this point, subjects will receive encouragement
and acknowledgement messages that will be spaced out
gradually until six months of abstinence. The patient
can contact the chat-bot at any time, so the number and
total time of interactions are a priori unknown and will
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 5 of 10

6. be recorded as secondary variables. In terms of proactive
contacts, the patient can choose their timing and
frequency.
Control group
Patients assigned to the control group will receive the
usual practice for smoking cessation, namely the differ-
ent interventions by primary care nurses and physicians
included in the services portfolio of the Madrid Regional
Health Service (Service 415, Care for adult smokers) that
consist of one or two visits prior to the quit date, follow-
up visits after one week and one month, and subsequent
control visits depending on whether the opportunity
arises, the professional’s criteria, and the patient’s needs.
Variables
The collaborating healthcare professionals will deliver
their data prior to the beginning of the study. The
recruiting participants will record patient data and will
be also responsible for their follow-up. All the informa-
tion will be recorded on a form created for the trial
(DCN). Each collaborator will access the form from their
personal computer via the project’s website, using a code
for personal identification. Two visits have been defined
for data collection of patients: basal and at 6 months.
Main outcome variable
Continuous tobacco abstinence at 6 months (yes/no). In
accordance to the recommendations by “the Russell
Standard” [20], abstinence is confirmed when the patient
declares to have smoked five or less cigarettes since the
beginning of the abstinence period and after being vali-
dated with a co-oxymeter Pico+TM Smokerlyzer® that
reads < 10 carbon monoxide particles per million (ppm)
in exhaled air.
Secondary outcome variables
Quality of life of the patient measured via the EuroQol-
5D-5 L, a generic questionnaire in their Spanish version
validated for our setting [21] that comprises two parts.
The first part consists of five questions on the health
condition of subjects to explore five dimensions: mobil-
ity, self-care, usual activities, pain/discomfort, and anx-
iety/depression. Each dimension is measured on a 5-
level scale. From these five questions, a single weighed
score is obtained, the so-called utility index. The higher
the score the better the health condition. To obtain this
index, the algorithm proposed for Spain has been
employed [22]. The second part consists of a visual
analog scale (VAS) ranging from 0 (worst possible con-
dition) to 100 (best possible health condition). The inter-
group differences in the utility score at 6 months will be
used to construct the quality-adjusted life years (QALYs)
gained with the intervention.
Employed time
Number of consultations and total time (in minutes) will
be recorded in the DCN by the healthcare professional
and automatically registered from the chat-bot.
Cost-utility
Costs derived from the intervention will be calculated
via a proxy of their cost-opportunity. The price of
avoided visits, calculated from the average salary of the
health worker’s professional category, will be categorized
as “saved costs”. An uncertainty threshold will be esti-
mated for all the applied costs comprising the time
range attributable to each activity. The calculation of
benefits will be done based on the QALYs, quantifying
the total gain from the EuroQol-5D-5 L utilities in the
intervention group at 6 months and assuming they will
stay constant for one year.
Adherence to treatment
Number of weeks the patient gets pharmacological treat-
ment prescribed by their doctor or nurse, as recorded by
the professional in the DCN during each visit or auto-
matically registered by the chat-bot.
Explanative and adjusting variables
Patient-related socio-demographic variables (first level):
1) age (years); 2) gender (M/W); 3) socio-economic
status (monthly income expressed as a multiple of the
minimum wage); 4) educational level (no education/pri-
mary education/secondary education/university educa-
tion); 5) nationality; 6) number of cigarettes/day; 7)
number of previous attempts to cease smoking; 8) basal
co-oximetry in exhaled air (ppm); 9) quit date; 10) level
of nicotine dependence (brief form of Fageström Test);
11) continuous cannabis use (defined as at least once
per week for the last month).
Healthcare centers and healthcare professionals (second
level): 1) gender (M/W); 2) age (years); 3) professional cat-
egory (medicine, nursing); 4) graduation year; 5) average
workload (measured as the average daily consultations per
health worker during the year prior to the trial).
Statistical analysis
The results will be presented in accordance with the
guidelines by the CONSORT declaration [23].
Description of basal characteristic
A descriptive analysis of the demographic and basal char-
acteristics of subjects in both groups will be conducted:
quantitative variables will be described by their measures
of central tendency (mean or median in the case of asym-
metric distributions) and their dispersion (standard devi-
ation or interquartile range for normal or non-normal
distributions, respectively). Qualitative variables will be
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 6 of 10

7. described by their proportions with their confidence inter-
vals set at 95% (CI 95%).
Basal comparison
Intergroup differences will be analyzed using statistical
tests for independent samples (Student’s t-test or chi-
squared test), whereas a test for related samples (re-
peated measures ANOVA) will be employed for differ-
ences within each group and across visits.
Main effectiveness analysis
Intergroup differences in the continuous abstinence
rates at 6 months (main variable) will be analyzed with
their CI 95%. Bivariate analysis: a Student’s t-test, or the
Mann-Whitney U-test in case the hypothesis of data
normality is rejected, will be used to compare the
outcome between groups in the case of quantitative
variables. For qualitative variables, either Pearson’s chi-
squared test or Fisher’s exact test, when applicable, will
be used. After adjusting for possible confounding factors,
the odds ratio of the association will be calculated using
a logistic regression model where the dependent variable
will be the continuous abstinence (yes/no) and the inde-
pendent variables will be patient-related variables and
the study group. In all cases, statistical significance will
be set at p < 0.05. Gained QALYs will be estimated at a
population level with their relevant CI 95% via paramet-
ric analysis methods and bootstrap techniques. Cost-
utility ratio: this aim is exploratory (there is no specific
design for it). The cost-utility ratio will be calculated
dividing the overall cost by the sum of potential gains
expressed in QALYs. For normally distributed data, a
multivariate sensitivity analysis will be conducted, mak-
ing the costs fluctuate within the constructed range of
uncertainty. For the same distribution, we will also make
the benefits (QALYs) fluctuate within the CI 95%. In all
cases, the analysis will be for intention-to-treat. The
number of withdrawals, losses to follow-up, and subjects
not complying with the protocol will be recorded together
with the relevant reasons. Losses to follow-up will be con-
sidered as smokers in the analysis [24]. A participant will
be considered lost to follow-up after missing two appoint-
ments, failing to be contacted by phone in at least two
occasions, and having no registered explanation for it.
Missing information on professionals and/or patients will
be replaced using values from the baseline data.
Discussion
Both the healthcare professionals accepting to partici-
pate in a clinical trial and the recruited patients may be
more interested in the subject of the trial, sharing an in-
creased awareness and motivation for the studied health
issues, which can lead to a bias. Although this can result
in better outcomes compared to trials with less selective
recruiting processes, it would have a conservative effect
(namely, it would decrease the magnitude of the inter-
group difference).
Additionally, participating collaborators may modify or
improve their prescribing habits due to their awareness of
being under observation (known as the Hawthorne effect).
This may reduce differences between the intervention and
control groups in terms of suitability of prescription.
A design by clusters was not chosen in this intervention,
since the possible influence among smokers in the same
healthcare center would be residual given that the app is
personal and with a restricted access for the patient. This
circumvents other biases resulting from variability in the
clinical practice of included professionals and their basal
knowledge on the subject of the intervention.
The intervention cannot be masked, so a blinded
evaluation of the main outcome variable has been de-
signed. The researchers in charge of the data analysis
and interpretation will be unaware of the allocation arm
of patients.
Although the assessment of the effect of the interven-
tion on quality of life is not confirmatory and the expected
changes are limited given the short period of time of the
trial, we consider it essential to include this outcome vari-
able in the analysis since intergroup differences can be
detected even if they are not strongly significant. Add-
itionally, measuring changes in quality of life allows for
incorporating outcome variables that are self-reported by
the patients, which facilitates the calculation of utilities in
the cost-utility analysis. This will enable to overcome
some deficiencies and limitations observed in previous in-
terventions by Patterson et al. [24].
The chat-bot can potentially increase the effectiveness
of the usual intervention for several reasons: on the one
hand, it will provide primary care professionals with a
very brief advice that will direct them to offer help for
cessation [10], which has proven its high effectiveness
for generating attempts to quit; on the other hand, it will
at the same time reduce their workload, increasing
their motivation to intervene on their patients. Add-
itionally, the overall time employed in the dishabitua-
tion process would significantly increase, a variable
that has been identified as one of the main predictors
for continuous abstinence [2]; lastly, reminding and
monitoring the use of pharmacological treatment will
predictably improve adherence to it and, therefore,
abstinence rates.
We estimate that this intervention will also show im-
proved accessibility, since it enables people to access an
effective program for quitting smoking even when phys-
ical limitations, schedule conflicts, or proximity issues
exist that hinder face-to-face consultations.
On the other hand, our pragmatic design [25] will
allow testing the usage of the chat-bot in conditions that
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 7 of 10

8. closely mimic real life. Both the recruitment of patients
and training of healthcare professionals, as well as the
information given to patients and characteristics of
healthcare centers, were the same in both arms. Other
aspects denoting its pragmatic design were the criteria
for inclusion, flexibility in applying the intervention,
comparison with usual practice, absence of formal visits
set by protocol, and intention-to-treat analysis.
It is essential to provide evidence on interventions di-
rected to outpatients who smoke and that, owing to new
information technologies, require less intervention by
healthcare professionals. Given the spread of smartphones
worldwide, there has been a great boom of apps that facili-
tate information on health issues and disease control and
treatment. However, new approaches based on informa-
tion and communications technologies appear at a much
higher speed than controlled, high-quality trials.
The prediction is that this trend will increase due to
the expanding number of users and the redundancy of
different approaches to manage one’s healthcare on ac-
count of the new available functions of technological ad-
vances, so it is important to study why this is happening
and remedy it in order to prevent ethical dilemmas and
harm to patients. In order to help identify effective inter-
ventions, we strongly believe that clinical trials must be
performed to evaluate new technologies and their related
approaches to help smokers quit, just as they are per-
formed for pharmacological drugs.
Abbreviations
App: Mobile Application; CI: Confidence interval; CO: Carbon Monoxide;
DCN: Data Collection Notebook; ERDF: European Regional Development
Fund; FIIBAP: Fundación para la Investigación e Innovación Biomédica en
Atención Primaria; FIS: Fondo de Investigaciones Sanitarias; NHS: British
National Health System; PIN: Personal Identification Number; PPM: Particles
Per Million (in exhaled air); QALYs: Quality-Adjusted Life Years; RR: Risk ratio;
SMS: Short Message Service. Text messages; VAS: Visual Analog Scale
Acknowledgements
We would like to thank the Research Support Unit from the Primary
Healthcare Management of the Community of Madrid regional government,
the Fundación para la Investigación e Innovación Biomédica en Atención
Primaria (FIIBAP) team, each of the many personnel of the 38 participating
health centers and Bumpho team.
Dejal@ Group (Contributor group) consists of the following researchers:
CS Ramón y Cajal (Alcorcón-Madrid): Garcia-Mesa, AM; Gomez, C; Liebana-
Nistal, E; Molina-Alameda, L.
CS Castillos (Alcorcón-Madrid): Andrade-Rosa, C, Caballero-Jiménez, A;
Escribano-Romo, G; Espejo-Romero, M; Ferrero-Brenes, A; Hernandez-Garcia,
JL; Perez-Perez, E; Zambrano-Alvarez, J.
CS Gregorio Marañon (Alcorcón-Madrid): Ayuso-Hernandez, E; Cuesta-
Monedero, MJ; Sanchez-Jimenez, FJ; Vigil-Escalera, A.
CS Ribota (Alcorcón-Madrid): Casado-Nistal, S; Garcia-Alva, M; Martin-Gomez,
I; Puyo-Rodriguez, N; Rodriguez-Cosio, A; Sanchez-Vazquez, I.
CS Dr Trueta (Alcorcón-Madrid): Alcala-Olmo, MA; Garcia-Alegre, MT; Garcia-
Alva, M; Hurtado-Monterrubio, C; Martin-Burillo, T; Moral-Moraleda, C.
CS Dr Luengo (Móstoles-Madrid): Ciria Holgado, E; Hermoso-Duran, C;
Hernandez-Rodriguez, S; Lafraya-Puente, AL; Melero-Serrano, JL; Molina-
Sanchez, MJ; Muñiz-Dominguez, E; Saiz-Loma-Osorio, C.
CS Alcalde Bartolomé (Móstoles-Madrid): Aparicio-Velasco, J; Cerezo-
Pascual, E; Fonte-Elizondo, L; Gomez-Martinez, MI; Marin-Osorio, C; Orozco-
Diaz, P; Prieto-Villegas, E; Rodriguez-Donoso, J; Rodriguez-Huete, A; Romero-
Blanco, FJ; Urbano-Fernandez, JP.
CS Felipe II (Móstoles-Madrid): Castillo-Portales, SJ; Collados-Navas, R;
Gutierrez-Sordo, P; Martin-Sacristan-Martin, MB; Sanchez-Quejido, M.
CS Barcelona (Móstoles-Madrid): Abad-Schilling, C; Arenas-Gonzalez, S;
Calonge-Garcia, ME; Capitan-Jurado, M; Menendez-Alvarez, S; Regalado-
Valle, MA.
CS Francia (Fuenlabrada-Madrid): Alache-Zuñiga, HC; Cerrada-Cerrada, E;
Garcia-Camps, R; Gutierrez-Notario, MJ; Mendez-Cabezas-Velazquez, J; Molina-
Paris, J; Pensado-Freire, H; Viñas-Fernandez, G.
CS Castilla La Nueva (Fuenlabrada-Madrid): Jaenes-Barrios, B; Jurado-Sueiro,
M; Ortiz-Sanchez, M; Santiago-Hernando, ML.
CS Cuzco (Fuenlabrada-Madrid): Arjona-Perez, S; Delgado-Mellado, I; Diaz-
Martin, R; Fandiño-Garcia, B; Gomez-Ortiz, MC; Lopez-Martin-Aragon, MT;
Martinez-Alvaro, C; Miguel-Abanto, MA; Noguerol-Alvarez, M; Parra-Roman, S;
Pascual-Garcia, Z; Ranz-Granados, AL; Rey-Rodriguez, S; Sanz-Perez, C.
CS Parque Loranca (Fuenlabrada-Madrid): Aguilar-Hurtado, E; Aguilera-
Rubio, M; Carbonero-Martin, AI; Esteban-Garcia, M; Esteban-Peña, C;
Fernandez-Sanchez, M; Garcia-Contreras, C; Hernandez-Sanchez, AM;
Herrera-Garcia, ML; Herrero-Fuentes, A; Oria-Rodriguez, M; Parra-Martin,
MS; Rojas-Giraldo, MJ; Saiz-Ordoño, A; Sanchez-Fernandez, YB; Villasevil-
Robledo, P.
CS Panaderas (Fuenlabrada-Madrid): Alonso-Ovies, M; Diaz-Martin, A;
Garcia-Arpa, N; Garcia-Rodriguez, F; Gonzalez-Carpio-Paredes, O; Grandes-
Muñoz, MJ; Jara-Peñacoba, M; Lopez-Villalbilla, A; Parellada-Ruiz-Cuesta, MJ;
Rabadan-Velasco, AI; Rodriguez-Bastida, M; Sobrado-Cemeño, C; Valdez-
Jaquez, A; Vicente-Perez, S.
CS Humanes de Madrid (Humanes de Madrid-Madrid): Matin-Porras, B.
CS Los Fresnos (Torrejón de Ardoz-Madrid): Gonzalez-Valls, A; Joglar-
Alcubilla, V; Santa-Cruz-Hernandez, J; Serrano-Serrano, ME.
CS Las Fronteras (Torrejón de Ardoz-Madrid): Castilla-Alvarez, C; Garcia-
Abad-Fernandez, MA; Garcia-Viada, M; Hernandez-Lachehab, S; Lopez-
Kollmer, L; Miguel-Ballano, A; Molina-Barcena, V; Ortega-Pineda, R; Sepulveda-
Gomez, I.
CS Las Veredillas (Torrejón de Ardoz-Madrid): Cuadrado-Gonzalez, P.
CS Luis Vives (Alcalá de Henares-Madrid):Jimenez-Moreno, MJ; Linares-
Sanchez, C; Pardo-Garcia, MA; Sanchez-Martin, F; Simon-Gutierrez, R; Solans-Aisa, B.
CS Juan de Austria (Alcalá de Henares-Madrid): Francisco-Romanillos, T;
Garcia-Ortega, A; Hernandez-Garcia-Alcala, P; Huguet-Vivas, F; Juarez-
Zapatero, ML; Lopez-Garcia, C; Manzano-Martin, MJ; Martinez-Torres, JA;
Miguel-Garzon, M; Rodrigo-Rodrigo, MP; Ruano-Dominguez, MA; Serna-Urnicia, A.
CS Reyes Magos (Alcalá de Henares-Madrid): Altares-Arriola, N; Aranzo-
Pacheco, R; Casado-Rodriguez, C; Cascao-Moutinho-Pereira, C; Escudero-Araus,
M; Fuentes-Manrique, C; Guijarro-Abanades, S; Hombrados-Gonzalo, P; Lopez-
Carabaño, AM; Meiriño-Perez, ML; Moneva-Vicente, GD; Noguera-Martinez, I;
Perez-Fernandez, M; Robres-Olite, M; Rubio-Rubio, T; Venegas-Gato, MF.
CS Miguel de Cervantes (Alcalá de Henares-Madrid): Bernal-Hertfelder, F;
Blas-Escribano, M; Castro-Fouz, MM; Castro-Sanchez, B; Dominguez-Perez, L;
Garcia-Gomez, PA; Herrero-Dios, A; Lor-Leandro, M; Lozano-Martin, I; Pacho-
Pinto, S; Pastor-Sanchez, R.
CS Los Hueros (Villalbilla-Madrid): Moreno-Chaparro, MD.
CS Villalbilla (Villalbilla-Madrid): Yagüe-Fernandez, E.
CS Dr. Cirajas (Madrid): Lopez, MJ; Mateo-Madruga, A; Palancar-Torre, JL;
Pumar-Sainz, P.
CS Guayaba (Madrid): Aragon-Marente, C; Gomez-Medina, MA; Granados-
Garrido, JA; Gutierrez-Sanchez, I; Heras-Alonso, MJ; Lopez-Gomez, MJ; Martin-
Peces, B; Martinez-Suberbiola, FJ; Medran-Gomez, A; Nuevo-Rodriguez, G;
Ochoa-Vilor, S; Vargas-Machuca, C.
CS Perales del Río (Perales del Rio-Madrid): Fernandez-Montes-Lopez-Mor-
ato, O; Jimenez-Rojas, R; Marcos-Frutos, C; Minguet-Arenas, C; Minue-Lorenzo,
C; Ruiz-Pascual, V; Sanchez-Fonseca, I; Sanchez-Gonzalez, JM.
CS Sta. Isabel (Leganés-Madrid): Alba-Gomez, F; Avila-Tomas, JF; Cayuela-
Mate, A; Cidoncha-Calderon, E; Cifuentes-Muñoz, AB; Fernandez-Garcia, RM;
Gala-Paniagua, JL; Gongora-Marin, A; Herrero-Municio, P; Moreno-Chocano-
Garcia-Carpintero, E; Perez-Cuadrado, S; Redondo-Horcajo, A; Vicente-
Sanchez, C; Villena-Romero, RM.
CS Leganés Norte (Leganés-Madrid): Hakami-Hakami, O.
CS María Jesús Hereza (Leganés-Madrid): Alvarez-Villalba, MM; Bedoya-
Frutos, MJ; Camarero-Palacios, J; Escobar-Gallegos, M; Hormigos-Agraz, A;
Innerarity-Martínez, J; Lopez-Lopez, MT; Perez-Ballesto, B; Perez-Gutierrez, E;
Tardaguila-Lobato, MP.; Terron-Barbosa, R.
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 8 of 10

9. CS El Greco (Getafe-Madrid): Ballarin-Gonzalez, A; Fernandez-Fernandez, Y;
Ferrer-Zapata, I; Gomez-Suarez, E; Lamonaca-Guasch, M; Morales-Ortiz, F;
Pelaez-Laguno, C; Quintana-Gomez, JL; Rioja-Delgado, E.
CS Sector III (Getafe-Madrid): Aza-Pascual, JI; Fuente-Arriaran, D; Garcia-
Perez, RA; Garriz-Aguirre, A; Gomez-Diaz, S; Gonzalez-Palacios, P; Gonzalez-
Sanchez, N; Peralta-Alvarez, G; Sanz-Velasco, C; Vazquez-Gallego, J; Vazquez-
Garcia, C.
CS El Restón (Valdemoro-Madrid): Fernandez-Duran, C; Peña-Anton, N.
CS Valle de la Oliva (Majadahonda-Madrid): Blanco-Canseco, JM; Gamez-
Cabero, I; Minguela-Puras, E; Serrano-Vega, J.
Study status
At the time of publication of this protocol, the study is at the data collection
stage.
Authors’ contributions
JFAT, EOE and CML conceived the study and participated in its design and
coordination. EEM participated in its design. FJMS, BMP and MESS
participated in different phases of the protocol design and coordination.
JFAT, EOE and CML wrote the manuscript, and FJMS, BMP and MESS
collaborated in the writing of the manuscript. All authors from the Dej@lo
group have read and approved the final manuscript.
Funding
This trial was funded by the Fondo de Investigaciones Sanitarias (FIS) del
Instituto de Salud Carlos III (ISCIII), the European Regional Development Fund
(ERDF) under registration number PI17/01942. The funders had no role in the
study design or in the decision to submit the report for publication. Also,
they will not take part in the collection or analysis of data, nor in the
assessment of the outcomes. The funding body has already performed a
peer-review of the trial protocol.
The main researcher received a grant from the Fundación para la
Investigación e Innovación Biomédica en Atención Primaria (FIIBAP) 2018 Call
for grants for promoting research programs.
Availability of data and materials
The datasets used and/or analyzed during the current study are available
from the corresponding author on reasonable request.
Ethics approval and consent to participates
This clinical trial was approved by the Clinical Research Ethics Committees of
the Community of Madrid (December 13th, 2017) and the University Hospital
12 de Octubre (Madrid, January 30th, 2018). Both committees are comprised
of qualified medical and scientific staff. The rights and wellbeing of patients
will be respected at all times. All patients will be properly informed both
verbally and in writing about the nature, objectives, risks, and potential
benefits of the trial. Informed consent, written, signed, and dated, will be
required from each subject.
The trial will comply with the Spanish law on human experimentation and
will respect all the ethical principles of autonomy, justice, goodness of
intention, and absence of intentionality in accordance with the good clinical
practice standards of the Declaration of Helsinki (Seoul, 2008) and the
Convenio de Oviedo (Oviedo, 1997).
A personal identification number (PIN) will be automatically generated that
will be the unique identifier of the patient, hence dissociating their clinical
and personal data. This PIN, together with a password, will also allow
subjects in the intervention group to access the chat-bot.
The chat-bot will not be used to collect any data allowing for identifying the
user.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Author details
1
Universidad Rey Juan Carlos, Madrid, Spain. 2
Healthcare center Sta. Isabel,
Madrid Regional Health Service, 28911, Leganés, Madrid, Spain. 3
Healthcare
center Los Castillos, DAO, Madrid Regional Health Service, Alcorcón, Madrid,
Spain. 4
Healthcare center Perales del Río, DAC, Madrid Regional Health
Service, Getafe, Madrid, Spain. 5
Healthcare center Guayaba, DAC, Madrid
Regional Health Service, Madrid, Spain. 6
Healthcare center Panaderas, DAO,
Madrid Regional Health Service, Fuenlabrada, Madrid, Spain. 7
Healthcare
center Los Fresnos, DAE, Madrid Regional Health Service, Torrejón de Ardoz,
Madrid, Spain. 8
Gerencia Asistencial de Atención Primaria (GAAP), Madrid,
Spain. 9
Health Services Research on Chronic Patients Network (REDISSEC),
Madrid, Spain.
Received: 17 October 2019 Accepted: 8 November 2019
References
1. Mathers C, World Health Organization. WHO global report mortality
attributable to tobacco [internet]. Geneva: World Health Organization; 2012.
[cited 2013 Jun 13]
2. Aveyard P, West R. Managing smoking cessation. BMJ. 2007;335:37–41.
3. National Institute for Health and Clinical Excellence. Smoking cessation
services in primary care, pharmacies, local authorities and workplaces,
particularly for manual working groups, pregnant women and hard to reach
communities. NICE public health guidance 10. London: NICE; 2008.
4. Treating tobacco use and dependence: 2008 update - NCBI Bookshelf
[Internet]. [cited 2012 Sep 28].
5. Minué-Lorenzo C, Olano-Espinosa E, del Cura-González I, Vizcaíno-Sánchez J,
Camarelles-Guillem F, Granados-Garrido J, et al. Subsidized pharmacological
treatment for smoking cessation by the Spanish public health system: a
randomized, pragmatic,clinical trial by clusters. Tob Induc Dis. 2019;17:64
[cited 2019, Sep 9];17(September).
6. Thorndike AN, Regan S, Rigotti NA. The treatment of smoking by US
physicians during ambulatory visits: 1994–2003. Am J Public Health. 2007;97:
1878–83.
7. Brotons C, Bjorkelund C. BulcM, et al. Prevention and health promotion in
clinical practice: the views of general practitioners in Europe. Prev Med.
2005;40(5):595–601.
8. Saito A. Health proffessionals perceptions of and potential barriers to
smoking cessation care. BMC Res Notes. 2010;3:329.
9. Pisinger C, Vestbo J, Borch-Johnsen K, Thomsen T, Jørgensen T. Acceptance
of the smoking cessation intervention in a large population-based study:
the Inter99 study. Scand J Public Health. 2005;33(2):138–45.
10. Aveyard P, Begh R, Parsons A, West R. Brief opportunistic smoking cessation
interventions: a systematic review and meta-analysis to compare advice to
quit and offer of assistance: Brief interventions for smoking cessation.
Addiction. 2012;107(6):1066–73.
11. The State of Broadband:Broadband catalyzing sustainable development.
International Telecommunications Union. Septiembre 2018.
12. Milani R, Franklin C. The Role of Technology in Healthy Living Medicine.
Prog Cardiovasc Dis. 2017;59:487–91.
13. Free C, Phillips G, Galli L, Watson L, Felix L, Edwards P, et al. The
effectiveness of mobile-health technology-based health behaviour change
or disease management interventions for health care consumers: a
systematic review. PLoS Med. 2013;10(1):e1001362.
14. Whittaker R, McRobbie H, Bullen C, Rodgers A, Gu Y. Mobile phone-based
interventions for smoking cessation. Cochrane Database Syst Rev. 2016;4:
CD006611.
15. Abroms LC, Westmaas JL, Bontemps-Jones J, Ramani R, Mellerson J. A
content analysis of popular smartphone apps for smoking cessation. Am J
Prev Med. 2013;45:732–6.
16. Haskins B, Lesperance D, Gibbons P, Boudreaux E. A systematic review of
smartphone applications for smoking cessation. Transl Behav Med. 2017;7:
292–9.
17. Aplicación de algoritmos genéticos y sistemas expertos en medicina
asistencial. Aplicaciones clínicas de la inteligencia artificial. Informes de
evaluación de Tecnologías sanitarias. AETSA 2009 / 6. ISBN: 978–84–
96990-87-6. Ed. Ministerio de Ciencia e Innovación, 2011. Available only
in Spanish.
18. Crutzen RY, Peters GJ, Portugal SD, Fisser EM, Grolleman J. An artificially
intelligent chat agent that answers adolescents’ questions related to sex,
drugs, and alcohol: an exploratory study. J Adolesc Health. 2011;48:514–9.
19. Grolleman J, Van Dijk B, Nijholt A, et al. Break the habit! Designing an e-
therapy intervention using a virtual coach in aid of smoking cessation. Lect
Notes Comput Sci. 2006;3962:133–41.
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 9 of 10

10. 20. West Rm Hayek P, Stead L, Stapleton J. Outcome criteria in smoking
cessation trials: proposal for a common standard. Addiction. 2005;100(3):
299–303.
21. Van Hout B, Janssen MF, Feng YS, et al. Interim scoring for the EQ-5D-
5L: mapping the EQ-5D-5L to EQ-5D-3L value sets. Value Health. 2012;
15(5):708–15.
22. Otero Lopez MJ, Codina JC, Tames Alonso MJ, et al. Medication errors:
standarizing the terminology and taxomany. Ruiz Jarabo 2000 grand results.
Farm Hosp. 2003;27(3):137–49.
23. Moher D, Hopewell S, Schulz KF, Montori V, Gøtzsche PC, Devereaux PJ,
Elbourne D, Egger M, Altman DG. CONSORT 2010 explanation and
elaboration: updated guidelines for reporting parallel group randomised
trials. BMJ. 2010;340:c869.
24. Barnes SA, Larsen MD, Schroeder D, Hanson A, Decker PA. Missing data
assumptions and methods in a smoking cessation study. Addiction. 2010;
105(3):431–7.
25. Thorpe KE, Zwarenstein M, Oxman AD, Treweek S, Furberg CD, Altman DG,
et al. A pragmatic-explanatory continuum indicator summary (PRECIS): a
tool to help trial designers. J Clin Epidemiol. 2009;62:464–75.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Avila-Tomas et al. BMC Medical Informatics and Decision Making (2019) 19:249 Page 10 of 10